Uncommon "soft" bolted joint 2

[eli28]

Hello everybody,
I have a bolted joint that is not common - one of the clamped members is made of ULTEM 1000.
Here is an illustration on the simplified clamped joint I tried to analyze:

In most cases we tend to use a joint in which its stiffness (Km) is larger by far than the bolt's stiffness (Kb).
It leads to the situation that in service bolt "feels" mainly the initial clamping force (Preload), Fi, and a small portion of the external load. This portion is determined by the relative stiffness Coefficient C=Kb/(Kb+Km) by the relation P_bolt=C*P, While P is the external exerted load and P_bolt is the relative part the bolt carries. In the typical cases of clamping steel members, it equals to C~0.8, and hence the entire force the bolt "feels" in service is Fi+C*P.
In my case the mutual stiffness of the clamped members is rather low thanks to the the ULTEM 1000 plate.
The preload in my case is 1500N and the external load P=780N.
I calculated and found that C~0.9, what means that the bolt is the main carrier of the external force, P, and that the clamped members should not "feel" a meaningful change in the initial clamping force, Fi, through Service.
I found that the bolt axial tensile force Fb=2200N and the clamped plates compression force Fm=-1430N.
I also checked in the literature that it makes sense and here you can see the right graph in which it shows the case like mine in which there is a "soft joint":

I was quite sure I am OK and predict fine what might be, but when I used the simulation to approve my prediction I was surprised to see that the opposite happened and that the members are the ones to take most of the external load like in the traditional case of clamping stiff steel plates.
What can you say about it?! Can someone verify in his own simulation tool what might happen?
Here is how I described the simplified problem in the simulation:
*pay attention that the upper (ULTEM 1000) member has a uniformly distributed force – a total of 780N.

 

[3DDave]

eli28,

Care to share a copy of the several analyses that show this result? A screen capture showing all the factors that you have suggested apply to your case?

It seems odd that a plate material that is only 1/60th of the modulus would have anywhere near 9X the stiffness of the bolt as shown in all the diagrams.

Then we can compare Kbolt, Kplate, Kjoint.

 

[rb1957]

if the original concern was crushing the ULTEM, use a spacer (so you don't clamp on the ULTEM.

another day in paradise, or is paradise one day closer ?

 

[eli28]

hey 3DDave - I am attaching a report of the analysis. please look at it and say what you think.

I marked with a red pen the places that may be relevant.
The photos are little blurred, but let me write down the important data:
Preload=1500 N
External Force=780 N
Fb=1580.6 N -> External force taken by the bolt = 1580.6-1500 = 80.6 N.
Fm=801N -> External force taken by the members = 1500-801 = 699 N.
I can't send the zip file with the analysis because it was run in my office's computer and I have no access to it.

rb1957 - there will be a washer to distribute the load.
regarding a spacer, is it necessary if I control the preload?

 

[3DDave]

Gotta say, that's the worst digital copy I've ever seen.

Without displacements there is no telling what the stiffness of any element is. We already knew the forces, so this adds no information.

I'd reduce the plates to 10mmx10mm and rerun it. It should make no difference to the typical load sharing calc unless the simulation is assuming the full square of material, far from the fastener, is compressed by the bolt.

It's good to do a sanity check like this when the results seem so abnormal.

 

[eli28]

 

[3DDave]

Where is the hole?

 

[eli28]

Bad photo...it's there.
Never mind, nothing good will be achieved from this hard copy
Do you have a simulation program and can check please what results you get?

 

[rb1957]

a washer helps distribute the load over more area of the ULTEM but I would still use a spacer (so as not to crush the ULTEM) ... unless I specified a very low torque (finger tight ?) then I'd worry about the nut shaking off ... maybe loktite on the nut after installation (this is clearly not a highly loaded part) ? wouldn't use a lock nut as the torque to overcome the lock is probably enough to crush the ULTEM.

another day in paradise, or is paradise one day closer ?

 

[rb1957]

why are you bolting ULTEM to stainless steel ??

another day in paradise, or is paradise one day closer ?

 

[eli28]

rb1957, maybe a spacer can be a good backup and protection from mistakenly overloading. The spacer aim is limiting the compression displacement the ULTEM will undergo?

I don't have to use ULTEM, but it should be a plastic part. Do you have any other suggestion to a stiff and high strength plastic that is easily machined?

 

[rb1957]

ideally the spacer has clearance over the ULTEM so it isn't being compressed at all. It could be flush, and being stiffer would prevent the ULTEN from being crushed.

Still not sure why you're bolting plastic to stainless steel.

another day in paradise, or is paradise one day closer ?

 

[BrianE22]

I don't have to use ULTEM, but it should be a plastic part.

Are you looking for an electrical insulator? I used Mica years ago when I was looking for a strong, machinable insulator. I don't know its modulus.

 

[eli28]

It isn't for insulation.
It's a structural element that must be lightweight, stiff, strong and shouldn't be made of metal for several reasons. What is the most rigid and strong plastic you know (besides composite that isn't an option)...?

 

[rb1957]

"stiff, light weight, strong, non-metallic" = graphite

another day in paradise, or is paradise one day closer ?

 

[3DDave]

Going to the hand-calcs I get:
the tension in the bolt should be around 2100N with the 1500N preload and the 780N external load.
the C for the joint is around 0.8

Ref:

https://mechanicalc.com/reference/bolted-joint-analysis#joint-stiffness

I did cheat and did not use the frustum formula; just tried a variety of calcs as if the plate was a uniformly compressed spacer. A range of approximations for the conical approximation showed only limited sensitivity.

The modulus of the plastic is low - the ratio of the areas is not.

 

[rb1957]

well, I would add a spacer and not compress the plastic ... that just feels "wrong".

another day in paradise, or is paradise one day closer ?

 

[eli28]

OK, here is an update - I got the help of Bolt Science calculator.
I am still confused and can't understand how most of the load is taken by the members.
You can see that the out of 780N the joint bears 712N. I have no explanation how it is reasonable.

 

[rb1957]

I couldn't see what you specified for the plates (1 SS, 1 plastic) ?

another day in paradise, or is paradise one day closer ?

 

[eli28]

YES

 

[goutam_freelance]

Your 'fixed part' plate is confusing. At which place the plate is fixed? It appears likely that it will be fixed at some edges. But where is the location?

Engineers, think what we have done to the environment !

https://www.linkedin.com/in/goutam-das-59743b30/